Introduction to R II
Shaurav Regmi
April 17, 2018
#Loading the packages and data as shown on the previous guide "Introduction to R".
packages.to.load <- c("plyr", "dplyr" ,"DT", "ggplot2", "plotly", "RODBC")
invisible(lapply(packages.to.load, library, character.only=TRUE))
load("C:/Users/shregmi/Documents/Current Projects/R Training/Example Datasets/Auto.rda")
write.csv(Auto, file="Autodata.csv")
auto.data <- read.csv(file ="C:/Users/shregmi/Documents/Current Projects/R Training/Autodata.csv", header = TRUE)1 Manipulating Data
This section will cover a couple ways to shape data into something that is easy to work with.
1.1 Using the $ operator
For very simple filters and selection of certain columns from a data frame, you can use the $ operator. Here, we take one column from our auto.data data frame and assign it to a new object. You can use the class() function to see what kind of object it is.
JustMPG <- matrix(auto.data$mpg)
class(JustMPG)## [1] "matrix"We can also pull multiple columns pretty easily.
MPGandWeight <- data.frame(auto.data$mpg, auto.data$weight)
head(MPGandWeight)## auto.data.mpg auto.data.weight
## 1 18 3504
## 2 15 3693
## 3 18 3436
## 4 16 3433
## 5 17 3449
## 6 15 43411.2 dplyr
The package dplyr is a heavily used library for data cleaning and preparation (one of R’s main strong suits). Here is a little cheat sheet covering some dplyr and tidyr features: https://www.rstudio.com/wp-content/uploads/2015/02/data-wrangling-cheatsheet.pdf
There are 5 main functions or “verbs” used in dplyr: Select, Filter, Mutate, Arrange, and Summarize. Many of these align with the main SQL commands: Select, Where, Group By, etc…
1.2.1 Select
Let’s take a subset of columns from the auto.data dataset.
auto.subset <- select(auto.data, name, mpg, cylinders, weight)
head(auto.subset)## name mpg cylinders weight
## 1 chevrolet chevelle malibu 18 8 3504
## 2 buick skylark 320 15 8 3693
## 3 plymouth satellite 18 8 3436
## 4 amc rebel sst 16 8 3433
## 5 ford torino 17 8 3449
## 6 ford galaxie 500 15 8 4341We can also specify which columns to exclude using the “-” operator.
auto.subset2 <- select(auto.data, -horsepower, -year)
head(auto.subset2)## X mpg cylinders displacement weight acceleration origin
## 1 1 18 8 307 3504 12.0 1
## 2 2 15 8 350 3693 11.5 1
## 3 3 18 8 318 3436 11.0 1
## 4 4 16 8 304 3433 12.0 1
## 5 5 17 8 302 3449 10.5 1
## 6 6 15 8 429 4341 10.0 1
## name
## 1 chevrolet chevelle malibu
## 2 buick skylark 320
## 3 plymouth satellite
## 4 amc rebel sst
## 5 ford torino
## 6 ford galaxie 5001.2.2 Filter
After selecting your data of interest, you can pass filters in the same way as you would a “where” clause in SQL.
only.4cyl <- filter(auto.subset, cylinders==4)## Warning: package 'bindrcpp' was built under R version 3.3.3head(only.4cyl)## name mpg cylinders weight
## 1 toyota corona mark ii 24 4 2372
## 2 datsun pl510 27 4 2130
## 3 volkswagen 1131 deluxe sedan 26 4 1835
## 4 peugeot 504 25 4 2672
## 5 audi 100 ls 24 4 2430
## 6 saab 99e 25 4 2375You can add multiple arguments to the filter function.
multi.filter <- filter(auto.subset, cylinders %in% c(4,6), weight<3000)
head(multi.filter)## name mpg cylinders weight
## 1 toyota corona mark ii 24 4 2372
## 2 plymouth duster 22 6 2833
## 3 amc hornet 18 6 2774
## 4 ford maverick 21 6 2587
## 5 datsun pl510 27 4 2130
## 6 volkswagen 1131 deluxe sedan 26 4 1835We can filter strings as well.
filter.name <- filter(auto.subset, cylinders==4, weight<2000, grepl("toyota|volkswagen", name))
head(filter.name)## name mpg cylinders weight
## 1 volkswagen 1131 deluxe sedan 26 4 1835
## 2 toyota corolla 1200 31 4 1773
## 3 volkswagen model 111 27 4 1834
## 4 volkswagen super beetle 26 4 1950
## 5 toyota corolla 1200 32 4 1836
## 6 volkswagen dasher 26 4 19631.2.3 Mutate
Using mutate, we can create new columns. Additionally, we can chain the previous “verbs” together using the “%>%” operator (called pipe operator). Note that the dplyr package must be loaded in order to use this operator.
add.col <- auto.data %>%
select(name, mpg, cylinders, weight) %>%
filter(weight>1800) %>%
mutate(weightpercyl <- weight/cylinders)
head(add.col)## name mpg cylinders weight
## 1 chevrolet chevelle malibu 18 8 3504
## 2 buick skylark 320 15 8 3693
## 3 plymouth satellite 18 8 3436
## 4 amc rebel sst 16 8 3433
## 5 ford torino 17 8 3449
## 6 ford galaxie 500 15 8 4341
## weightpercyl <- weight/cylinders
## 1 438.000
## 2 461.625
## 3 429.500
## 4 429.125
## 5 431.125
## 6 542.6251.2.4 Arrange
This is equivalent to an “order by” statement in SQL. Here, we sort the previous output by “weightpercyl” in descending order using the “-” operator.
add.col.sorted <- auto.data %>%
select(name, mpg, cylinders, weight) %>%
filter(weight>1800) %>%
mutate(weightpercyl = weight/cylinders) %>%
arrange(-weightpercyl)
head(add.col.sorted)## name mpg cylinders weight weightpercyl
## 1 mazda rx-4 21.5 3 2720 906.6667
## 2 peugeot 504 19.0 4 3270 817.5000
## 3 mercedes-benz 240d 30.0 4 3250 812.5000
## 4 peugeot 505s turbo diesel 28.1 4 3230 807.5000
## 5 mazda rx-7 gs 23.7 3 2420 806.6667
## 6 peugeot 504 27.2 4 3190 797.50001.2.5 Summarize
summarized.auto <- auto.data %>%
select(name, mpg, cylinders, weight) %>%
filter(weight>1800) %>%
mutate(weightpercyl = weight/cylinders) %>%
summarise(avg_mpg = mean(mpg),
min_weight = min(weight),
median_weightpercyl = median(weightpercyl))
summarized.auto## avg_mpg min_weight median_weightpercyl
## 1 23.18903 1825 544.1667We can pass a group_by() clause as well.
summarized.auto2 <- auto.data %>%
select(name, mpg, cylinders, weight) %>%
filter(weight>1800) %>%
mutate(weightpercyl = weight/cylinders) %>%
group_by(cylinders) %>%
summarise(avg_mpg = mean(mpg),
avg_weight = mean(weight),
avg_weightpercyl = mean(weightpercyl))
summarized.auto2## # A tibble: 5 x 4
## cylinders avg_mpg avg_weight avg_weightpercyl
## <int> <dbl> <dbl> <dbl>
## 1 3 20.6 2398. 800.
## 2 4 29.0 2331. 583.
## 3 5 27.4 3103. 621.
## 4 6 20.0 3202. 534.
## 5 8 15.0 4115. 514.2 Visualization
This section will cover a few useful packages and methods for visualization in an R Markdown document.
2.1 Displaying Tables in R Markdown
When using print() or head() functions to show the contents of a data frame, the output looks very ugly. There are a huge number of packages that are specifically for displaying tables in an aesthetically pleasing way. The “DT” or Data Tables package is one of them.
datatable(auto.data)There are a huge number of arguments you can add to make your reports look nicer and to add functionality. BONUS: You’ll also write the smallest piece of JavaScript code in this as well.
datatable(auto.data, extensions = 'Buttons', options = list(
dom = 'Bfrtip',
buttons= c('copy', 'excel'),
pageLength = 5,
initComplete = JS("
function(settings, json) {
$(this.api().table().header()).css({
'background-color': '#232D69',
'color': '#fff'
});
}")
), rownames = FALSE)2.2 Charts with Plotly
Cheat sheet for plotly: https://images.plot.ly/plotly-documentation/images/r_cheat_sheet.pdf.
Another great resource: https://plot.ly/r/
plot <- plot_ly(x = auto.data$weight, type="histogram")
plotIn plotly, you can add multiple sets of traces or bars to the same plot with the use of pipe operators. You can also name each of the traces/bars so that the result is clear.
plot2 <- plot_ly(data =auto.data, alpha = 0.5) %>% #adjusting color transparency
add_histogram(x = auto.data$weight[auto.data$origin == 1], name="American") %>%
add_histogram(x = auto.data$weight[auto.data$origin == 2], name="European") %>%
add_histogram(x = auto.data$weight[auto.data$origin == 3], name="Japanese") %>%
layout(barmode = "overlay")
plot2Here is an example of a scatterplot generated by plotly. This has many more features and much more functionality than the plot we first generated with the R Base plot() function.
auto.data$OriginName[which(auto.data$origin == 1)] = "American"
auto.data$OriginName[which(auto.data$origin == 2)] = "European"
auto.data$OriginName[which(auto.data$origin == 3)] = "Japanese"
plot3 <- plot_ly(data = auto.data, x = ~weight, y = ~mpg, color = ~OriginName, type ="scatter", mode="markers", text= ~paste("Car name: ", name ,
"</br> Year: ", year,
"</br> Cylinders: ", cylinders))
plot32.3 Charts with ggplot2
Cheat sheet for ggplot2: https://www.rstudio.com/wp-content/uploads/2015/03/ggplot2-cheatsheet.pdf
